Performance Optimization Of Layered Bed Pressure Swing Adsorption For Hydrogen Purification

Hydrogen is the excellent source of energy in the universe,which is ubiquitous in nature and industrial waste gas.As an excellent carrier of clean and renewable energy,hydrogen has the potential to replace traditional fossil fuels.Hydrogen energy has a wide range of applications.It can not only be used as a common raw material for chemical and pharmaceutical enterprises,but also as a fuel source for hydrogen fuel cell vehicles.However,the large-scale application of hydrogen energy is still limited by the production,purification,storage and transportation and so on.Among them,hydrogen purification is of great significance to the safety,efficiency,reliability and durability of the hydrogen utilization system.As a kind of gas separation technology,Pressure Swing Adsorption?PSA?is one of the important methods to remove impurities from mixed gas and purify hydrogen.At present,many adsorption equilibrium theories have been proposed,which can be used for equation fitting of adsorption isotherms under different experimental conditions.When the adsorption isotherms constructed by these basic theories?such as Langmuir adsorption equilibrium theory?are directly used to predict the adsorption amount of gas under different experimental conditions,the results are not so good.Therefore,in order to build a suitable adsorption isotherm model,the Sips isotherm model,Toth isotherm model,extended Langmuir isotherm model,extended Langmuir-Freundlich isotherm model and Virial Equation model are successively introduced based on the Langmuir adsorption equilibrium theory.The adsorption isotherms of single components?CO2,CO,N2,CH4,Ar,H2?in activated carbon and zeolite LiX adsorption beds are studied.The adsorption isotherm data are measured at 293 K,308 K and 323 K,respectively,and within the pressure range of 0¹.0 MPa.The experimental data are compared with the simulation results of Langmuir,Sips and Toth isotherm models.Then,according to the adsorption isotherm data provided in the references,the artificial neural network?ANN?model is trained and the adsorption isotherm is predicted.The adsorption isotherms of H₂/CO/CH₄ on activated carbon adsorption bed and zeolite adsorption bed are studied.A breakthrough curve in adsorption is the course of the effluent adsorptive concentration at the outlet of a fixed bed adsorber.Breakthrough curves are important for adsorptive separation technologies and for the characterization of porous materials.Layered bed has been widely used in the PSA purification process of hydrogen.Compared with the adsorption bed with a single adsorbent,the layered bed has better hydrogen purification performance.In this thesis,the breakthrough curves of the layered bed based on activated carbon/zeolite 5A is studied.In order to obtain the optimal data for predicting the breakthrough curves by using the artificial neural network,the PSA cycle model is run in Aspen adsorption software.The breakthrough curve of PSA model is verified by experimental data in literature.Then a data set is obtained based on the verified PSA cycle model to train the ANN model used to predict the breakthrough time.In this work,a triple layered feed-forward back propagation neural network is built as a surrogate model to capture the nonlinearity in the data of PSA cycle performance,which is a kind of artificial neural network.Artificial neural network architecture consists of an input layer,an output layer and a hidden layer.After designing the network structure,the sample data set is used to train the neural network.Through the study of the results,it is confirmed whether the neural network model is suitable for the analysis and prediction of PSA cycle performance.In order to get a tradeoff of purity and recovery of hydrogen,a multi-objective optimization is performed based on the ANN model.The influences of multi-component gases,adsorption bed structure and PSA operation parameters are studied in this work.Based on the artificial neural network model and multi-objective optimization function,the optimization design criteria of hydrogen purification system are further proposed,which are of great theoretical significance and practical value for improving PSA purification performance.